This is a test. let me know what you think.
## -- Attaching packages --------------------------------------- tidyverse 1.3.1 --## v ggplot2 3.3.5 v purrr 0.3.4
## v tibble 3.1.2 v dplyr 1.0.7
## v tidyr 1.1.3 v stringr 1.4.0
## v readr 1.4.0 v forcats 0.5.1## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()## [1] "ReadMe" "Raw_Data" "Clean_Data" "Cleaning_Log" "Deletions"## New names:
## * enrolled_boys_6_11...258 -> enrolled_boys_6_11...259
## * enrolled_boys_6_11...333 -> enrolled_boys_6_11...335I am checking your cleaning log against your clean data.
## Joining, by = "binding"dd$check_clean %>% table(useNA = "ifany")## .
## FALSE TRUE <NA>
## 4376 11885 9921Out of your 25k entries, you have 12k correct, 4k not correct and 10k as NA.
Let’s now just focus on the NA
dd %>%
filter(is.na(check_clean)) %>%
group_by(Question) %>%
tally() %>%
arrange(desc(n))dd %>% head()Question: from your cleaning log
uuid: uuid
Old Value: from your cleaning log
New value: from your cleaning log
Reason: from your cleaning log
value_raw : value from raw dataset
value_clean: value from clean dataset
check_clean: New Value (from cleaning log) is the same as value_clean (from clean data)
logg %>% filter(Question == "All") %>% mutate(uuid_dup = duplicated(uuid)) there are 1681 survey deleted, there are 1676 in the deletion log.
there are 9 uuid duplicated in the cleaning log for deletion
action: none needed.
dd %>% filter(Question == "sanitation_features_other", is.na(check_clean)) it seems there are 10% of the household who does not have toilets.
action: to be check with the skip logic
dd %>% filter(Question == "main_source_water_other", is.na(check_clean)) dd %>%
filter(Question == "main_source_water_other", is.na(check_clean)) %>%
select(uuid, `New Value`) %>%
left_join(select(cleann, main_source_water, uuid)) ## Joining, by = "uuid"action: Explore what happened, was the “other” recoded?
dd %>% filter(Question == "when_arrived_current_location", is.na(check_clean)) %>%
View()In your cleaning log, the “New Value” is empty, while there is a value in the clean dataset. No reason written
action: Explore what happened
dd %>% filter(Question == "when_leave_place_origin", is.na(check_clean)) %>%
View()Same as above.
dd %>% filter(Question == "67", is.na(check_clean)) Not sure what 67 variable is for, but same as above. In addition, all of them were turn to 67 while the value were different
Action: investigate on your side.
dd %>% filter(Question == "sanitation_facilities_problems_other", is.na(check_clean)) %>%
pull(Reason) %>% table()## .
## Clarification from enumerator
## 339
## Clarification from enumerator + Translating to English
## 16
## other options recoded
## 4
## other options recoded in the choice list
## 22
## Translating to English
## 272dd %>% filter(Question == "sanitation_facilities_problems_other", is.na(check_clean),
Reason != "other options recoded",
Reason != "other options recoded in the choice list") dd %>%
filter(Question == "sanitation_facilities_problems_other", is.na(check_clean),
Reason != "other options recoded",
Reason != "other options recoded in the choice list") %>%
select(uuid, `New Value`) %>%
left_join(select(cleann, sanitation_facilities_problems, uuid)) %>%
select(`New Value`, sanitation_facilities_problems) %>% table()## Joining, by = "uuid"## sanitation_facilities_problems
## New Value latrines_too_far other other
## Big problems 0 2
## Dont have sanitation facilities 0 4
## Dont have toilet 0 216
## Dont have toilet, need to dig and build a toilet 0 2
## Dont have toilet, share with a family 0 2
## Go to open spaces 0 2
## Go to open spaces' 0 2
## Goto op[en spaces 0 2
## No , dont have toilet 0 16
## No sanitation facilities 0 4
## None of the above 0 2
## Only have and open bit 0 2
## open defication 0 1
## out toilet needs renovation 0 2
## There are no toilets 0 2
## There is no sanitation facility 0 22
## Use with neighbour 0 2
## We dont have toilet cleaning materials 0 0
## We go out at night 0 0
## we used open deficaiton 0 19
## Yes 0 2
## sanitation_facilities_problems
## New Value other facilities_too_crowded
## Big problems 0
## Dont have sanitation facilities 0
## Dont have toilet 2
## Dont have toilet, need to dig and build a toilet 0
## Dont have toilet, share with a family 0
## Go to open spaces 0
## Go to open spaces' 0
## Goto op[en spaces 0
## No , dont have toilet 0
## No sanitation facilities 0
## None of the above 0
## Only have and open bit 0
## open defication 0
## out toilet needs renovation 0
## There are no toilets 0
## There is no sanitation facility 0
## Use with neighbour 0
## We dont have toilet cleaning materials 0
## We go out at night 2
## we used open deficaiton 0
## Yes 0
## sanitation_facilities_problems
## New Value unclean_unhygienic other
## Big problems 0
## Dont have sanitation facilities 0
## Dont have toilet 0
## Dont have toilet, need to dig and build a toilet 0
## Dont have toilet, share with a family 0
## Go to open spaces 0
## Go to open spaces' 0
## Goto op[en spaces 0
## No , dont have toilet 0
## No sanitation facilities 0
## None of the above 0
## Only have and open bit 0
## open defication 0
## out toilet needs renovation 0
## There are no toilets 0
## There is no sanitation facility 0
## Use with neighbour 0
## We dont have toilet cleaning materials 2
## We go out at night 0
## we used open deficaiton 0
## Yes 0Same as for main source of water. Check if recoding happened or should happen. Action: explore what happened when recoding other.
dd %>%
filter(Question == "shelter_issues_other", is.na(check_clean)) dd %>%
filter(Question == "shelter_issues_other", is.na(check_clean))%>%
select(uuid, `New Value`) %>%
left_join(select(cleann, shelter_issues, uuid)) %>%
select(`New Value`, shelter_issues) %>% table()## Joining, by = "uuid"## shelter_issues
## New Value lack_bathing
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 4
## None 0
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0
## shelter_issues
## New Value lack_bathing lack_cooking lack_lights_inside lack_lights_outside
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 0
## None 2
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0
## shelter_issues
## New Value lack_bathing lack_cooking lack_lights_outside
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 2
## None 0
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0
## shelter_issues
## New Value lack_cooking lack_cooking other lack_lights_outside
## Its nice 0 0 0
## Its normal 0 0 0
## My house is safe 0 0 0
## no issue 0 0 0
## No problem 0 0 0
## No problems 0 0 0
## No, No problem 0 0 2
## None 2 0 0
## None of the above 0 0 0
## Nothing 0 0 0
## Thanks to Allah 0 0 0
## There is no issue 0 0 0
## There is no problem 0 0 0
## shelter_issues
## New Value lack_lights_outside lack_lights_inside
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 2
## None 0
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0
## shelter_issues
## New Value lack_lights_outside unsafe_bathing lack_privacy
## Its nice 0 0
## Its normal 0 0
## My house is safe 0 0
## no issue 0 0
## No problem 0 0
## No problems 0 0
## No, No problem 2 0
## None 0 2
## None of the above 0 0
## Nothing 0 0
## Thanks to Allah 0 0
## There is no issue 0 0
## There is no problem 0 0
## shelter_issues
## New Value lack_space lack_privacy lack_lights_outside lack_lights_inside unsafe_cooking
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 2
## None 0
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0
## shelter_issues
## New Value other other lack_lights_inside unsafe_bathing
## Its nice 2 0 0
## Its normal 2 0 0
## My house is safe 2 0 0
## no issue 0 0 0
## No problem 8 0 0
## No problems 6 0 0
## No, No problem 98 0 2
## None 52 0 0
## None of the above 2 0 0
## Nothing 2 0 0
## Thanks to Allah 4 0 0
## There is no issue 2 0 0
## There is no problem 32 0 0
## shelter_issues
## New Value unsafe_bathing other
## Its nice 0
## Its normal 0
## My house is safe 0
## no issue 0
## No problem 0
## No problems 0
## No, No problem 2
## None 0
## None of the above 0
## Nothing 0
## Thanks to Allah 0
## There is no issue 0
## There is no problem 0It seems there is “no issues”. Should it be removed as you did then? Or removed? Action: investigate if re-coding correct.
dd %>%
filter(Question == "hh_main_source_income_other", is.na(check_clean)) %>%
select(uuid, `Old Value`) %>%
left_join(select(cleann, shelter_issues, uuid))## Joining, by = "uuid"dd %>%
filter(Question == "hh_main_source_income_other", is.na(check_clean)) %>%
pull(Reason) %>%
table() ## .
## Anwser was choice list Clarification from choices
## 98 177
## Clarification from enumerator Clarification from translation
## 1 48
## irrelevent entries other options recoded in the choices
## 58 10
## other options recoded inthe choices reclasified
## 8 11
## response already in the list of choices
## 3dd %>%
filter(Question == "hh_main_source_income_other", is.na(check_clean)) %>%
select(uuid, `Old Value`) %>%
left_join(select(cleann, starts_with("hh_main_source_income"), uuid)) ## Joining, by = "uuid"does Clarification from choices and Clarification from translation means re-classfied?
Action: Check if clarification were correclty re-coded/classified.
dd %>%
filter(Question == "common_type_ids_other", is.na(check_clean)) %>%
select(uuid, `Old Value`, `New Value`) %>%
left_join(select(cleann, starts_with("common_type_ids"), uuid)) ## Joining, by = "uuid"Same as the the others “other”
Action: check recoding and action.
For all others that are removed because of “none”, maybe because of missing skip logic during the coding, do you need to also remove all the questions? or add a “none” in your select mutiple? Or do you want just to report as that and include the “no-issues” in the denominator? e.g. do you want to report : (1) XX % of households reported to have an issue. From the ones reporting an issue, YY% reported as OPTION1 as an issue. (2) Amongst all the households, XX% reported no issues, YY% reported OPTION1 as an issue. for (1) you would want to recode the none as missing for all options so that you can use the skiplogic.
dd %>% filter(check_clean == F)dd %>% filter(check_clean == F) %>% nrow()## [1] 4376logg2 <- logg %>%
mutate(`New Value` = ifelse(`New Value` == "FALSE", 0, `New Value`),
`New Value`= ifelse(`New Value` == "TRUE", 1, `New Value`))
old_new_values2 <- mapply(old_new,
cleaning_log = split(logg2, row.names(logg2)),
variable = "Question",
MoreArgs = list(
data_raw = raww,
data_clean = cleann,
uuid_raw = "uuid",
uuid_clean = "uuid",
uuid_cleaning_log = "uuid"
),
SIMPLIFY = F) %>% do.call(rbind, .)
dd2 <- logg2 %>%
mutate(binding = paste0(uuid, Question)) %>%
left_join(old_new_values2) %>%
mutate(check_raw = `Old Value` == value_raw,
check_clean = `New Value` == value_clean) %>%
select(-c(binding, ID, `Follow-up`, Enumerator, Community, `Modified by?`, Notes, check_raw),
uuid, Question,`Old Value`, `New Value`, value_raw, value_clean, check_clean, Reason)## Joining, by = "binding"dd2 %>% filter(check_clean == F) %>% nrow()## [1] 705With changing the T/F we are reducing to 705 wrong values!
dd2 %>% filter(check_clean == F)It seems "casuel _labour _Wages _construction _etc" has a typo.
dd2 %>% filter(check_clean == F,
`New Value` != "casuel _labour _Wages _construction _etc") %>% nrow()## [1] 535Down to 535.
dd2 %>% filter(check_clean == F,
`New Value` != "casuel _labour _Wages _construction _etc")It seems some values are off.
Action: please check those.
inspec <- cleaninginspectoR::inspect_all(cleann, "uuid")It seems you have lots of outliers.
inspec %>%
filter(issue_type %in% c("log normal distribution outlier", "normal distribution outlier")) %>%
group_by(variable) %>%
tally()Action: Please check all those variables for LOW and HIGH values.
I don’t have the KOBO so I am just checking the obvious one
cleann %>%
select(males_0_2y:total_hh) %>%
mutate(across(.fns = as.numeric)) %>%
mutate(total_cal = rowSums(across(males_0_2y:females_60_older)),
check = total_cal == total_hh) %>%
pull(check) %>%
table(useNA = "ifany")## .
## TRUE
## 11645All good
Action: none needed
cleann$age_pregnent_female_give_birth %>% table(useNA = "ifany")## .
## 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
## 3 21 2 2 8 7 23 51 42 99 36 72 61 68 182 59
## 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
## 78 120 85 210 53 95 49 33 172 28 40 50 20 43 12 8
## 43 44 45 46 48 50 55 <NA>
## 7 1 10 2 1 2 1 9789Action: change the 1 year old pregnant woman to NA?
I am not sure what does estimate_hh_income refers to but you have very low values all the estimate_hh_income.XXXX have very low values. People earning .20 dollars. See below for anything less than 5(USD?) (0’s excluded)
cleann %>%
select(uuid, starts_with("estimate_hh_income.")) %>%
mutate(across(starts_with("estimate_hh_income."), ~(. < 5 & . > 0), .names = "{.col}_less_5"))Action: needed please checked those values.
cleann %>%
mutate(across(.cols = males_0_2y:females_12_17y, .fns = as.numeric)) %>%
mutate(total_cal = rowSums(across(males_0_2y:females_12_17y)),
check = total_cal == hh_children) %>%
pull(check) %>%
table(useNA = "ifany")## .
## FALSE TRUE
## 32 11613cleann %>%
mutate(across(.cols = males_0_2y:females_12_17y, .fns = as.numeric)) %>%
mutate(total_cal = rowSums(across(males_0_2y:females_12_17y)),
check = total_cal == hh_children) %>%
filter(check == F) %>%
select(uuid, males_0_2y:females_12_17y,hh_children, total_cal)There are calculations that were not corrected. Action: Please ammend
cleann %>%
mutate(across(.cols = males_3_5y:females_12_17y, .fns = as.numeric)) %>%
mutate(total_cal = rowSums(across(males_3_5y:females_12_17y)),
check = total_cal == school_age_children) %>%
filter(check == F) %>%
select(uuid, males_3_5y:females_12_17y,school_age_children, total_cal)As above.
The KOBO calculation includes children males and females from 6 to 11. That variable is only used to filter on the pregnant_female_give_birth. If you are not going to analyse it, just delete it
After checking for only females from 12 to 59, you should turn about 380 pregnant_female_give_birth and the 2 follow-up questions pregnant_female_give_birth_yes, age_pregnent_female_give_birth to NA.
cleann %>%
mutate(across(.cols = c(females_12_17y, females_18_40, females_41_59), .fns = as.numeric)) %>%
mutate(check_female = rowSums(across(c(females_12_17y, females_18_40, females_41_59))) > 0) %>%
select(check_female, pregnant_female_give_birth) %>%
table(useNA = "ifany")## pregnant_female_give_birth
## check_female no prefer_not_answer yes <NA>
## FALSE 338 5 38 132
## TRUE 9261 53 1818 0cleann %>%
mutate(across(.cols = c(females_12_17y, females_18_40, females_41_59), .fns = as.numeric)) %>%
mutate(check_female = rowSums(across(c(females_12_17y, females_18_40, females_41_59))) > 0,
to_be_fixed = check_female == F & !is.na(pregnant_female_give_birth) ) %>%
filter(to_be_fixed == T) %>%
select(uuid, check_female, pregnant_female_give_birth, pregnant_female_give_birth_yes, age_pregnent_female_give_birth) You are suming \({seeing_memb}+\){hearing_memb}+\({walking_climbing_memb}+\){remembering_memb}+\({selft_care_memb}+\){communicating_memb} and then create a loop from it. If you a person had 2 difficulites, you would have a sum of 2, hence creating a loop for 2 persons and not one. Action: how are you going to deal with that?
cleann %>%
mutate(across(.cols = shelt_count.bedrooms:shelt_count.other_sh, .fns = as.numeric)) %>%
mutate(total_cal = rowSums(across(shelt_count.bedrooms:shelt_count.other_sh)),
check = total_cal == shelt_count.sum_rooms) %>%
filter(check == F) %>%
select(uuid, males_3_5y:females_12_17y,school_age_children, total_cal)There 1849 calculation to be corrected.
Action: please ammend or remove the column if not needed
From your script you shared, you have a list of checks that were not implemented:
- pregnant_female_give_birth_yes > 2
- age_pregnent_female_give_birth < 15
- children_working.permanent_job_child_male > 3
- children_working.permanent_job_child_female > 3
- sharing_sanitation_facilities_yes > 14 “how more then 14 hh share a latrine?”
- how_many_shelters > 8 “how one hh can own 8 or more shelters?”
- shelter_rooms_note > 3
- shelt_count.bedrooms > 3
- shelt_count.living_rooms > 3
- shelt_count.kitchens > 3
- shelt_count.toilets shelt_count.other_sh > 3
- shelt_count.sum_rooms > 3
Why were those not checked?
cleann %>%
select(uuid, shelt_count.bedrooms:shelt_count.other_sh) %>%
mutate(across(shelt_count.bedrooms:shelt_count.other_sh, ~ (.> 3), .names = "{.col}_above_3")) %>%
summarise(across(ends_with("above_3"), sum))cleann %>%
select(uuid, shelt_count.bedrooms:shelt_count.other_sh) %>%
mutate(across(shelt_count.bedrooms:shelt_count.other_sh, ~ (.> 3), .names = "{.col}_above_3"))I am not sure what this refers to but the number of months is not the conversion of years into months. Please check again those variables as they currently do not make sense.
cleann %>%
mutate(hhs1 = case_when(how_often_fs_happen == "often_10_plus" ~ 2,
how_often_fs_happen %in% c("rately_1_2", "sometimes_3_10") ~ 1,
is.na(how_often_fs_happen) ~ 0),
hhs2 = case_when(often_go_sleep_night_hungry == "often_10_plus" ~ 2,
often_go_sleep_night_hungry %in% c("rately_1_2", "sometimes_3_10") ~ 1,
is.na(often_go_sleep_night_hungry) ~ 0),
hhs3 = case_when(often_not_eating_whole_day_night == "often_10_plus" ~ 2,
often_not_eating_whole_day_night %in% c("rately_1_2", "sometimes_3_10") ~ 1,
is.na(often_not_eating_whole_day_night) ~ 0),
hhs_score = hhs1 + hhs2 + hhs3,
hhs_cat = cut(hhs_score, breaks = c(0,5,7), labels = c("low-none-severe", "extreme"), right = F)) %>%
select(hhs_cat, district) %>% table(useNA = "ifany") %>% prop.table(2)## district
## hhs_cat abdulaziz aden_yabaal afgooye afmadow baardheere
## low-none-severe 1.000000000 1.000000000 1.000000000 1.000000000 1.000000000
## extreme 0.000000000 0.000000000 0.000000000 0.000000000 0.000000000
## district
## hhs_cat badhaadhe badhan baidoa baki balcad
## low-none-severe 1.000000000 1.000000000 0.996316759 1.000000000 0.989071038
## extreme 0.000000000 0.000000000 0.003683241 0.000000000 0.010928962
## district
## hhs_cat bandarbayla baraawe belet_weyne belet_xaawo berbera
## low-none-severe 1.000000000 1.000000000 0.985542169 1.000000000 0.994871795
## extreme 0.000000000 0.000000000 0.014457831 0.000000000 0.005128205
## district
## hhs_cat boondheere borama bossaso bulo_burto burco
## low-none-severe 0.935064935 0.984536082 0.988165680 0.971014493 1.000000000
## extreme 0.064935065 0.015463918 0.011834320 0.028985507 0.000000000
## district
## hhs_cat burtinle buuhoodle buur_hakaba cabudwaaq cadaado
## low-none-severe 0.993975904 1.000000000 0.991452991 0.993103448 1.000000000
## extreme 0.006024096 0.000000000 0.008547009 0.006896552 0.000000000
## district
## hhs_cat cadale caluula caynabo ceel_afweyn ceel_barde
## low-none-severe 1.000000000 1.000000000 0.970760234 0.993750000 0.992366412
## extreme 0.000000000 0.000000000 0.029239766 0.006250000 0.007633588
## district
## hhs_cat ceel_buur ceel_dheer ceel_waaq ceerigaabo daynile
## low-none-severe 1.000000000 1.000000000 1.000000000 0.994623656 0.982248521
## extreme 0.000000000 0.000000000 0.000000000 0.005376344 0.017751479
## district
## hhs_cat dharkenley dhuusamarreeb diinsoor doolow eyl
## low-none-severe 0.982352941 1.000000000 1.000000000 1.000000000 1.000000000
## extreme 0.017647059 0.000000000 0.000000000 0.000000000 0.000000000
## district
## hhs_cat gaalkacyo_north gaalkacyo_south galdogob garbahaarey
## low-none-severe 1.000000000 1.000000000 1.000000000 1.000000000
## extreme 0.000000000 0.000000000 0.000000000 0.000000000
## district
## hhs_cat garowe gebiley hamar_jaab_jab hamar_weyne
## low-none-severe 0.992957746 0.995327103 1.000000000 0.933333333
## extreme 0.007042254 0.004672897 0.000000000 0.066666667
## district
## hhs_cat hargeysa hawl_wadaag heliwa hobyo hodan
## low-none-severe 1.000000000 0.978021978 1.000000000 0.992907801 0.989417989
## extreme 0.000000000 0.021978022 0.000000000 0.007092199 0.010582011
## district
## hhs_cat iskushuban jalalaqsi jamaame jariiban jowhar
## low-none-severe 1.000000000 1.000000000 1.000000000 1.000000000 0.983516484
## extreme 0.000000000 0.000000000 0.000000000 0.000000000 0.016483516
## district
## hhs_cat kahda karaan kismayo kurtunwaarey laas_caanood
## low-none-severe 0.984375000 0.975728155 1.000000000 1.000000000 0.984210526
## extreme 0.015625000 0.024271845 0.000000000 0.000000000 0.015789474
## district
## hhs_cat laasqoray lughaye luuq marka mataban
## low-none-severe 1.000000000 0.990783410 1.000000000 1.000000000 0.978417266
## extreme 0.000000000 0.009216590 0.000000000 0.000000000 0.021582734
## district
## hhs_cat owdweyne qandala qansax_dheere qardho qoryooley
## low-none-severe 0.992592593 1.000000000 1.000000000 0.988023952 1.000000000
## extreme 0.007407407 0.000000000 0.000000000 0.011976048 0.000000000
## district
## hhs_cat rab_dhuure sablaale shangaani sheikh shibis
## low-none-severe 1.000000000 1.000000000 1.000000000 1.000000000 1.000000000
## extreme 0.000000000 0.000000000 0.000000000 0.000000000 0.000000000
## district
## hhs_cat taleex tayeeglow waaberi waajid wadajir
## low-none-severe 1.000000000 1.000000000 1.000000000 1.000000000 0.945205479
## extreme 0.000000000 0.000000000 0.000000000 0.000000000 0.054794521
## district
## hhs_cat wanla_weyn wardhiigleey xudun xudur yaaqshiid
## low-none-severe 1.000000000 1.000000000 0.991869919 0.994382022 0.952941176
## extreme 0.000000000 0.000000000 0.008130081 0.005617978 0.047058824
## district
## hhs_cat zeylac
## low-none-severe 1.000000000
## extreme 0.000000000bonndheere, wadajir and yaarqshiid are at the 5% threshold. If any HHS score for a strata is above 5 % we should check with Matt and Katie.
Action: Are you going to report as Banadir as a whole or for each city boundary ? (maybe Khada and Dharkenley)